Authors: | Martinez-Ortiz, Carlos Žunić, Dragiša Žunić, Joviša |
Affiliations: | Mathematical Institute of the Serbian Academy of Sciences and Arts | Title: | Measuring shape rectangularities | First page: | 369 | Last page: | 374 | Conference: | 2011 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT) | Issue Date: | 2011 | Rank: | M33 | ISBN: | 9781467307529 | DOI: | 10.1109/ISSPIT.2011.6151590 | Abstract: | This paper introduces a family of rectangularity measures. Several rectangularity measures already exist in literature, and all of them evaluate how much the shape considered differs from a perfect rectangle. But all existing measures assume that all rectangles have the same shape, and a consequence is that these measures do not distinguish among the rectangles whose edge-ration differs. In this paper we consider that differently elongated rectangles have different shape and propose a family of new rectangularity measures which distinguish among such rectangles. New rectangularity measures are invariant with respect to translation, rotation and scaling transformations. In addition, we introduce a tuning parameter which enables an additional flexibility of the new measures with respect to performing applications. |
Keywords: | image processing | image retrieval | measuring shape | rectangularity | Shape | Publisher: | IEEE |
Show full item record
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.